If your stack speaks OpenAI, it already speaks Wordcab.
Wordcab's endpoints are OpenAI-compatible. Swap base_url, keep your existing SDK. Official Python and TypeScript clients, a single-binary CLI, and ready-made adapters for Pipecat, LiveKit, Vapi, and Retell.
SDKs, CLI, and compatible clients.
The full SDK reference lives in /docs/sdks; the CLI reference is at /docs/cli. This page is the index.
pip install wordcabPython SDK
Sync and async clients. Retries, pagination, streaming helpers. Pydantic models for structured outputs. Python 3.9+.
npm install @wordcab/sdkTypeScript SDK
ESM-first, tree-shakeable. Full types for streaming, tool use, JSON mode. Node 18+, Deno, Bun.
wordcab CLI
Scripts, CI, operators. Transcription, speech, agents, calls, deployments, logs. Homebrew, npm, or signed static binary.
Existing OpenAI code
Change two lines — base_url and api_key. Chat, embeddings, and audio endpoints are 1:1 compatible.
Bare HTTP
Same endpoints with just a bearer token. Shell scripts, pipeline smoke tests, and anything else that speaks curl.
Voice-agent frameworks
Drop Wordcab Voice and Think into the orchestration layer your team already runs. Full matrix on the integrations page.
Need SDK access or integration guidance now?
Building on Wordcab and need SDK access, integration examples, or implementation support before this page goes live? Talk to the team directly.
Talk to an Engineer
We usually respond within one business day.